Comparison of Different Approaches to the Creation of a Mathematical Model of Melt Temperature in an LD Converter

Research on prediction model of converter temperature and carbon content based on spectral feature extraction

The flame of converter mouth can well reflect the change of temperature and composition of molten steel in the furnace. The flame characteristics of converter mouth collected by device can well predict the smelting process of converter. Based on the flame spectrum data set of converter mouth, this paper uses the BEADS algorithm and rough set attribute reduction algorithm optimized by genetic algorithm to extract the features of 2048-dimensional wavelength data. Through the model, eight indexes that contribute greatly to temperature and carbon content are selected, which are f-507, f-520, f-839, f-1073, f-1371, f-1528, f-1727 and f-1826. The MIC coefficients of the eight indicators with temperature and carbon content are calculated, and the MIC coefficients of the variables is small, and the selected indicators are representative. There was a significant correlation between temperature and C content. In BP neural network of temperature prediction model, it is found that the prediction accuracy of the training set is 0.99, the prediction accuracy of the test set is 0.99, the prediction accuracy of the verification set is 0.99, and the prediction accuracy of the whole set is 0.99. Through statistics, it is found that the hit rate of the temperature model in the range of ± 5 K is 88.7%, and the hit rate in the range of ± 10 K is 98.4%. and the RMSE parameter analysis shows that the average prediction error is 3.85 K. In BP neural network of carbon content prediction model, it is found that the prediction accuracy of the training set is 0.99, the prediction accuracy of the test set is 0.99, the prediction accuracy of the verification set is 0.98, and the prediction accuracy of the whole set is 0.99. Through statistics, it is found that the hit rate of the carbon contents model in the range of ± 0.05% is 94.0%, and the hit rate in the range of ± 0.10% is 98.3%, and the RMSE parameter analysis shows that the average prediction error is 0.021%. Finally, the universality of the model is verified by MIV algorithm.

Prediction of Temperature and Carbon Concentration in Oxygen Steelmaking by Machine Learning: A Comparative Study

The basic oxygen steelmaking process (BOS) faces the issue of the absence of information about the melt temperature and the carbon concentration in the melt. Although deterministic models for predicting steelmaking process variables are being developed in metallurgical research, machine-learning models can model the nonlinearities of process variables and provide a good estimate of the target process variables. In this paper, five machine learning methods were applied to predict the temperature and carbon concentration in the melt at the endpoint of BOS. Multivariate adaptive regression splines (MARS), support-vector regression (SVR), neural network (NN), k-nearest neighbors (k-NN), and random forest (RF) methods were compared. Machine modeling was based on static and dynamic observations from many melts. In predicting from dynamic melting data, a method of pairing static and dynamic data to create a training set was proposed. In addition, this approach has been found to predict the dynamic behavior of temperature and carbon during melting. The results showed that the piecewise-cubic MARS model achieved the best prediction performance for temperature in testing on static and dynamic data. On the other hand, carbon predictions by machine models trained on joined static and dynamic data were more powerful. In the case of predictions from dynamic data, the best results were obtained by the k-NN-based model, i.e., carbon, and the piecewise-linear MARS model in the case of temperature. In contrast, the neural network recorded the lowest prediction performance in more tests.